The new western Lake Ontario and Niagara Peninsula seismicity patterns for the 1992-2001 period are significantly different from the patterns of the mainly historical earthquakes seen in the (1819-1969) period. The latter figure covered the period when instrumentation and number of recording stations were either non-existent or very poor. It is interesting to note that during the whole 150 year history from 1819 to 1969, there was only one earthquake detected under Lake Ontario and that one occurred under the southeastern part of the lake. There were no epicentres in the western Lake Ontario region. Unlike the more recent record of seismicity, all of the epicentres lie on the land areas both north and south of the lake with the greatest number occurring in the Niagara Peninsula. A cluster of earthquakes is apparent in both figures in the Hamilton area, however significant differences exist in the orientation and areal shape of this cluster. The magnitudes of the historical earthquakes were also, on average, larger than the ones shown in (1992-2001) period. The most plausible explanation for the major change in in patterns that during the time when the instrumentation was poor, most of the epicentres were located primarily on the basis of the "felt" reports of people living on the land areas around the lake. Furthermore, isoseismal intensity data were in general not available for small events. Substantial reductions in the location uncertainty have been achieved by the installation of the SOSN, and it appears that enough data have now accumulated to identify several well constrained spatial trends. The emerging patterns of seismicity are consistent with several areas of known seismicity (past and present). We agree with Stevens (1994) who concluded that discrepancies between seismicity maps from different eras can be attributed to differences in the location uncertainty for the earthquake epicentres.
The overall seismicity patterns from the SOSN data sets are generally consistent with the patterns in the (1970-1991) period. This is the data that were recorded after the great expansion in the number of seismic stations in the Canadian network. Our patterns show additional clusters of events confirming continued seismic activity exists in the Clarendon-Linden fault (Jacobi et al. 1993), the area just east of Buffalo , and southern shore of Lake Erie near the Ohio-Pennsylvania border (Seeber and Armbruster 1993, Marciera et al. 2000). In the latter paper fault plane solutions were presented for the events south of the Lakes including the recent Pymatuning earthquake . They indicate a "roughly east-west or north-south vertical strike-slip mechanism which agrees well with existing estimates of the stress field." The recent SOSN patterns does not show the cluster of events which was observed in the (1970-1991) map near Gobles. This cluster actually represents over 400 induced earthquakes with magnitudes 0 to 3.4 which resulted from pumping of fluids in oil recovery activities during the period 1980-1984 (Mereu et al. 1986). All of the earthquakes occurred over a 5 km x 5 km area over the oil field. The Figure shows these events to be oriented in a N-S line about 25 km long. This orientation and alignment is really an artifact of poor control in the N-S direction inherent in the GSC solutions because of their station distribution. After the pumping stopped the earthquakes stopped and thus were not observed in the 1992-2001 data.